Rethinking Table Pruning in TableQA: From Sequential Revisions to Gold Trajectory-Supervised Parallel Search

TL;DR

This paper introduces TabTrim, a novel parallel search framework for table pruning in TableQA that improves accuracy by aligning pruning with gold trajectories during training and inference.

Contribution

It transforms table pruning from sequential revisions to gold trajectory-supervised parallel search, enhancing accuracy in tabular reasoning tasks.

Findings

TabTrim achieves 73.5% average accuracy, outperforming baselines by 3.2%.

It reaches 79.4% on WikiTQ and 61.2% on TableBench.

State-of-the-art performance across diverse tabular reasoning tasks.

Abstract

Table Question Answering (TableQA) benefits significantly from table pruning, which extracts compact sub-tables by eliminating redundant cells to streamline downstream reasoning. However, existing pruning methods typically rely on sequential revisions driven by unreliable critique signals, often failing to detect the loss of answer-critical data. To address this limitation, we propose TabTrim, a novel table pruning framework which transforms table pruning from sequential revisions to gold trajectory-supervised parallel search. TabTrim derives a gold pruning trajectory using the intermediate sub-tables in the execution process of gold SQL queries, and trains a pruner and a verifier to make the step-wise pruning result align with the gold pruning trajectory. During inference, TabTrim performs parallel search to explore multiple candidate pruning trajectories and identify the optimal sub-table. Extensive experiments demonstrate that TabTrim achieves state-of-the-art performance across diverse tabular reasoning tasks: TabTrim-8B reaches 73.5% average accuracy, outperforming the strongest baseline by 3.2%, including 79.4% on WikiTQ and 61.2% on TableBench.